Board level semiconductor packaging and interconnection of integrated circuits (IC) is the process to electrically connect the IC chips to each other and to external circuitry to function as an electronic system. The IC chips have input and output contact pads and the interconnection is typically an array of metallic connections within a support substrate. The normal packaging process of forming a fully functional product based on integrated circuits usually includes placing the semiconductor IC chips on a printed circuit board (PCB) and soldering their contact pads to contact pads on the PCB. Several connection techniques are widely used and well known in the art. These include wire bonding, tape automated bonding (TAB), flip-chip bonding, etc.
The earliest process is wire bonding, the process of placing the ICs face up on the PCB, and bonding fine wire conductors from the IC contact pads to the PCB pads. Wire bonding is by far the most common and economical connection technique, usually by thermocompression, thermosonic or ultrasonic processes. Because wire bonding requires wires to be welded to the chip, there must be adequate space to accommodate the wires.
This technique is appropriate to the use of solid PCBs and rigid ICs, formed by dicing silicon wafers that are typically 0.5 mm thick. It is desirable for many applications to have flexible electronic products, for example flexible display backplanes, in which case the PCB must be replaced by a printed circuit tape (PCT), often known in the industry as a “flex circuit”. In addition, the integrated circuit must be flexible. Flexible integrated circuits can be fabricated by direct deposition and patterning of semiconductor and other materials to form interconnected transistors on plastic (or other flexible substrate). However, the performance of such devices tends to be less than optimal, and the processes for fabricating them less well developed and more costly than for the well-known processes of fabricating transistors in silicon wafers.
TAB utilizes patterned metal on a polymeric tape to join the chips together, involving bonding gold-bumped pads on the chips to external circuitry. TAB requires mechanical force such as pressure or a burst of ultrasonic vibration and elevated temperature to accomplish the welding between the wires or bumps and the designated surface.
One way to obtain higher performance flexible circuits at affordable cost is to thin conventional wafers until they are flexible. It is now common in the industry to produce wafers thinner than 100 microns, and as thin as 20 microns or less, by a combination of grinding, polishing, and etching, and these processes add a modest fraction to the cost of the ICs. In principle, these very thin wafers (which are now quite flexible) may be diced and mounted on flexible substrates, resulting in flexible products. In practice, many problems arise in the handling of such thin (and therefore delicate) chips.
One approach, a process called Fluidic Self-Assembly, attempts to address these issues by mechanically indenting a polymer substrate to approximately the thickness of a thinned silicon chip having contact pads, which then is inserted into the depression. However, apart from the possibility of defects arising from incomplete filling of the holes, this process does not result in a highly planar surface due to variations in the thickness of chips coupled with variations in the depth of the depressions, and there is still a small but significant gap between the side of the chip and the side of the hole. This leads to complexity in forming the subsequent interconnect metallization. It also does not address the issue of reliability due to mismatched thermal expansions.
Another possible approach to these problems is to apply, by lamination or casting followed by lithography, a thin film of polymer with cutouts that are the right size into which to insert the thinned ICs. This process could be low cost and high speed. However, as with the Fluidic Self-Assembly process, there would still be imperfections in the coplanarity due to the inevitable variations, and there would still be a gap between the sides of the chip and the side of the cutout.
The increase in density of input/output (I/O) lines caused the industry to shift to so-called “flip-chip” methods, in which the IC is placed on the PCB face down, and a direct metal-to-metal contact is made between the pads. Most commonly, this is some variant of a process in which a low-melting metal (solder) is first placed on the pads of one member (for example by dipping into a container of liquid metal, after the surface has been treated in such a way that solder will adhere only to the pads), forming solder “bumps”. The other member also has “bumps” formed on the pads; the non-solder bumps are some appropriate metal which must be part of the final lithographic processing step. The IC is placed on the PCB by a pick-and-place machine with enough accuracy to orient the solder bumps over the correct locations, and with sufficient heat to liquefy the solder, a strong metal-metal connection is then made. The empty space between connections is filled (“underfilled”) with epoxy so as to strengthen the adhesion and prevent failure due to corrosion of the metal and to mechanical stress from heating/cooling cycles and the mismatch of coefficients of thermal expansion (CTE) between the IC and the PCB.
A major advantage of flip-chip bonding over wire bonding and TAB is that the connection paths are shorter, and therefore have better electrical characteristics. In addition, flip-chip bonding requires minimal mounting area which results in further overall cost saving. However, while flip-chip technology has tremendous advantages over wire bonding and TAB, its cost and technical limitations are significant. For example, the cost of forming bumps on the pads is high, and the underfilling process of an adhesive between the chip and the support substrate increases both manufacturing complexity and cost. Furthermore, the solder joints exhibit increased electrical resistance as well as cracks and voids over time due to fatigue from thermo-mechanical stresses. Finally, the solder is typically high temperature, which is not suitable to flexible substrate.